In [1]:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from tensorflow import keras
import seaborn as sns
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout
from tensorflow.keras.models import load_model
In [2]:
#Loading the Historical fires data of Oregon state into a pandas dataframe
historical_fires = pd.read_csv('./data/historical_oregon_fires.csv', low_memory=False)
In [3]:
historical_fires.shape
Out[3]:
In [5]:
historical_fires.describe().T
Out[5]:
In [7]:
historical_fires.corr()
Out[7]:
In [ ]:
In [9]:
drop_columns = ['OBJECTID', 'Serial' ,'FireNumber' ,'FiscalYear','FireArea','Current_District', 'Current_Unit','FireName',
'PreparedBy', 'LandmarkLocation', 'Equipment','NumberPersonnel', 'CreationDate', 'ModifiedBy', 'ModifiedDate',
'Twn', 'Rng', 'Sec', 'Subdiv', 'County','LO_Name','DiscoveredByName','Specific',
'Cause_Comments','Lead_Investigator', 'Degree_certain','General_Restriction', 'Industrial_Restriction',
'RegulatedUseZone','Involve_op', 'Notif_year','Notif_dist', 'Notif_number','Report_DateTime', 'Attack_DateTime',
'Control_DateTime','Walk_In_Delay', 'DS_Cost', 'ES_Cost', 'PC_Cost', 'OA_Cost','TotalCost', 'Est_Damage', 'CostRecovery',
'AttackBy','AttackType','Size_class','Size_prot', 'Homes_Saved', 'Homes_Lost','Structures_Saved', 'Structures_Lost',
'Number_of_Injuries','Number_of_Deaths', 'CauseType']
In [10]:
historical_fires.drop(columns=drop_columns,inplace=True)
In [11]:
use_later = ["Burn_Index","WeatherStationID"]
historical_fires.drop(columns=use_later,inplace=True)
In [12]:
#Convert columns to datetime format
historical_fires['Ign_DateTime']=pd.to_datetime(historical_fires['Ign_DateTime'])
historical_fires['Discover_DateTime']=pd.to_datetime(historical_fires['Discover_DateTime'])
In [13]:
historical_fires['fire_dayofyear'] = historical_fires['Discover_DateTime'].dt.dayofyear
#historical_fires['fire_dayofyear'] = historical_fires['fire_dayofyear'].astype(int)
In [14]:
historical_fires = historical_fires.fillna(0)
In [15]:
#Removing time columns
time_columns = ['Ign_DateTime','Discover_DateTime']
historical_fires.drop(columns=time_columns,inplace=True)
In [18]:
#The target to predict, area burnt by fire.
historical_fires['Size_acres'].describe()
Out[18]:
In [19]:
#checking for Null Values in the data
historical_fires.isnull().sum()
Out[19]:
In [20]:
import matplotlib.pyplot as plt
import numpy as np
# Normal Distribution of fire size (target)
# Central Limit Theory applies
plt.hist(historical_fires['Size_acres'], bins=np.logspace(-4,6, 50));
plt.ylabel('Number of Fires')
plt.xlabel('Fire Size')
plt.gca().set_yscale("log")
plt.gca().set_xscale("log")
plt.show()
In [23]:
# !pip install plotly-express
import plotly.express as px
fig = px.scatter(historical_fires, x = "Longitude", y = "Latitude", color = "FireYear", opacity=0.2)
fig.show()
0, 0 is Null Island: https://en.wikipedia.org/wiki/Null_Island¶
Often when working with data that includes latitudes and longitudes, 0,0 is an identifier for potentially erroneous data
In [24]:
# 7200 observations at Null Island
len(historical_fires.query('Longitude==0 & Latitude==0'))
Out[24]:
In [25]:
# 99.99% of latitudes are between ...
np.percentile(historical_fires['Latitude'], 0.01), np.percentile(historical_fires['Latitude'], 99.99)
Out[25]:
In [26]:
# 99.99% of longitudes are between ...
np.percentile(historical_fires['Longitude'], 0.01), np.percentile(historical_fires['Longitude'], 99.99)
Out[26]:
Remove Outliers¶
In [27]:
# Remove all the Null Islands
historical_fires = historical_fires[(historical_fires['Latitude'] >0) &
(historical_fires['Longitude']<0)]
In [28]:
# Look at the relationship between latitude/longitude and year that the fire occurred in
fig = px.scatter(historical_fires, x = "Longitude", y = "Latitude", color = "FireYear", opacity=0.2)
fig.show()
In [29]:
# Look at the relationship between latitude, longitude, fire year and fire size (in acres)
fig = px.scatter(historical_fires, x = "Longitude", y = "Latitude", color = "FireYear", size="Size_acres", opacity =0.9)
fig.show()
In [30]:
# Overlaying fire latitude/longitude to stamen-terrain mapbox layout
fig = px.scatter_mapbox(historical_fires, lat='Latitude', lon='Longitude', color='FireYear', opacity=0.1)
fig.update_layout(mapbox_style='stamen-terrain')
fig.show()
In [31]:
# Cluster the locations
from sklearn.cluster import KMeans
kmeans = KMeans(n_clusters=10, n_jobs=-1)
historical_fires['cluster'] = kmeans.fit_predict(historical_fires[['Longitude', 'Latitude']])
px.scatter(historical_fires, x='Longitude', y='Latitude', color='cluster')
In [80]:
# Cluster the locations based on GIS acres
from sklearn.cluster import KMeans
kmeans = KMeans(n_clusters=7, n_jobs=-1)
explore = historical_fires.copy()
explore['cluster'] = kmeans.fit_predict(explore[['Size_acres']])
px.scatter(explore, x='Longitude', y='Latitude', color='cluster')
In [50]:
def engineer_features(df):
# What's the terrain, based on latitude & longitude?
clusters = pd.get_dummies(df['cluster'], prefix='cluster')
for col in clusters:
df[col] = clusters[col]
# Remove Outlier/invalid values for 'caused_by' column
df = df[df['Caused_by'] < 15]
#As the standard deviation is very high, removing the outliers from the data
df = df[df['Size_acres'] < 500]
#Based on pandas-profiling report, we're rejecting features highly-correlated with other criteria
correlated_columns = ['Slope', 'Aspect', 'Topography', 'X', 'Y']
df = df.drop(columns=correlated_columns)
# remove unusable variance - constant value
unusable_variance = ['FireCategory']
df = df.drop(columns=unusable_variance)
return df
In [51]:
# wrangle dataframe using engineering features function
historical_fires = engineer_features(historical_fires)
In [54]:
#Creating the data for training
Y = historical_fires['Size_acres']
X = historical_fires.drop(columns='Size_acres')
In [55]:
# #Normalising the Data
sc = StandardScaler()
X_norm = sc.fit_transform(X)
In [56]:
#creating the train and test data
X_train,X_test,y_train,y_test = train_test_split(X_norm,Y,test_size = 0.2)
print("# Train: {} , #Test: {}".format(X_train.shape[0], X_test.shape[0]))
print("# inputs: {}".format(X_train.shape[1]))
n = X_train.shape[1]
In [58]:
cmap = sns.cubehelix_palette(dark=.3, light=.8, as_cmap=True)
ax = sns.scatterplot(x="Longitude", y="Latitude",
hue="Caused_by", size="Size_acres",
sizes=(20, 200),palette="Set1", data=historical_fires)
In [41]:
import pandas_profiling
pandas_profiling.ProfileReport(X_train)
Out[41]:
Overview
Dataset info
| Number of variables | 42 |
|---|---|
| Number of observations | 41669 |
| Total Missing (%) | 0.0% |
| Total size in memory | 10.4 MiB |
| Average record size in memory | 262.0 B |
Variables types
| Numeric | 19 |
|---|---|
| Categorical | 0 |
| Boolean | 17 |
| Date | 0 |
| Text (Unique) | 0 |
| Rejected | 6 |
| Unsupported | 0 |
Warnings
Aspectis highly correlated withTopography(ρ = 0.99903) RejectedBehaviorhas 20800 / 49.9% zeros ZerosElevationis highly correlated withSlope(ρ = 0.99764) RejectedFederal_landshas 3469 / 8.3% zeros ZerosFireCategoryhas constant value 1 RejectedFlame_lengthhas 29522 / 70.8% zeros ZerosLand_Class_tghas 20772 / 49.9% zeros ZerosLatitudeis highly correlated withY(ρ = 0.99991) RejectedLongitudeis highly correlated withX(ρ = 0.99964) RejectedSB_360_Lothas 38352 / 92.0% zeros ZerosSize_at_attackis highly skewed (γ1 = 181.37) SkewedSize_at_attackhas 21569 / 51.8% zeros ZerosSlopeis highly correlated withAspect(ρ = 0.99941) Rejectedclusterhas 4815 / 11.6% zeros Zeros
Variables
Aspect
Highly correlated
This variable is highly correlated with Topography and should be ignored for analysis
| Correlation | 0.99903 |
|---|
Behavior
Numeric
| Distinct count | 9 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 1.0486 |
|---|---|
| Minimum | 0 |
| Maximum | 9 |
| Zeros (%) | 49.9% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 0 |
| Q1 | 0 |
| Median | 1 |
| Q3 | 2 |
| 95-th percentile | 3 |
| Maximum | 9 |
| Range | 9 |
| Interquartile range | 2 |
Descriptive statistics
| Standard deviation | 1.4522 |
|---|---|
| Coef of variation | 1.3849 |
| Kurtosis | 11.138 |
| Mean | 1.0486 |
| MAD | 1.0608 |
| Skewness | 2.6526 |
| Sum | 43695 |
| Variance | 2.1089 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 0.0 | 20800 | 49.9% | |
| 2.0 | 11689 | 28.1% | |
| 1.0 | 5973 | 14.3% | |
| 3.0 | 1786 | 4.3% | |
| 4.0 | 682 | 1.6% | |
| 9.0 | 601 | 1.4% | |
| 6.0 | 61 | 0.1% | |
| 7.0 | 49 | 0.1% | |
| 5.0 | 28 | 0.1% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 20800 | 49.9% | |
| 1.0 | 5973 | 14.3% | |
| 2.0 | 11689 | 28.1% | |
| 3.0 | 1786 | 4.3% | |
| 4.0 | 682 | 1.6% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 4.0 | 682 | 1.6% | |
| 5.0 | 28 | 0.1% | |
| 6.0 | 61 | 0.1% | |
| 7.0 | 49 | 0.1% | |
| 9.0 | 601 | 1.4% |
Caused_by
Numeric
| Distinct count | 15 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 4.9731 |
|---|---|
| Minimum | 0 |
| Maximum | 14 |
| Zeros (%) | 0.4% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 1 |
| Q1 | 1 |
| Median | 5 |
| Q3 | 8 |
| 95-th percentile | 13 |
| Maximum | 14 |
| Range | 14 |
| Interquartile range | 7 |
Descriptive statistics
| Standard deviation | 3.8403 |
|---|---|
| Coef of variation | 0.7722 |
| Kurtosis | -0.70066 |
| Mean | 4.9731 |
| MAD | 3.1701 |
| Skewness | 0.61847 |
| Sum | 207230 |
| Variance | 14.748 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 1.0 | 13559 | 32.5% | |
| 5.0 | 7390 | 17.7% | |
| 9.0 | 5026 | 12.1% | |
| 13.0 | 3293 | 7.9% | |
| 7.0 | 2432 | 5.8% | |
| 2.0 | 2255 | 5.4% | |
| 6.0 | 2070 | 5.0% | |
| 3.0 | 1824 | 4.4% | |
| 10.0 | 1296 | 3.1% | |
| 4.0 | 1100 | 2.6% | |
| Other values (5) | 1424 | 3.4% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 161 | 0.4% | |
| 1.0 | 13559 | 32.5% | |
| 2.0 | 2255 | 5.4% | |
| 3.0 | 1824 | 4.4% | |
| 4.0 | 1100 | 2.6% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 10.0 | 1296 | 3.1% | |
| 11.0 | 262 | 0.6% | |
| 12.0 | 113 | 0.3% | |
| 13.0 | 3293 | 7.9% | |
| 14.0 | 91 | 0.2% |
Discovered_By
Numeric
| Distinct count | 14 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 8.2852 |
|---|---|
| Minimum | 0 |
| Maximum | 99 |
| Zeros (%) | 0.4% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 1 |
| Q1 | 3 |
| Median | 6 |
| Q3 | 8 |
| 95-th percentile | 9 |
| Maximum | 99 |
| Range | 99 |
| Interquartile range | 5 |
Descriptive statistics
| Standard deviation | 16.266 |
|---|---|
| Coef of variation | 1.9632 |
| Kurtosis | 26.498 |
| Mean | 8.2852 |
| MAD | 5.6874 |
| Skewness | 5.2694 |
| Sum | 345230 |
| Variance | 264.57 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 6.0 | 12197 | 29.3% | |
| 8.0 | 6086 | 14.6% | |
| 1.0 | 3730 | 9.0% | |
| 3.0 | 3602 | 8.6% | |
| 9.0 | 3175 | 7.6% | |
| 7.0 | 3090 | 7.4% | |
| 2.0 | 2952 | 7.1% | |
| 4.0 | 2917 | 7.0% | |
| 5.0 | 1951 | 4.7% | |
| 99.0 | 1270 | 3.0% | |
| Other values (4) | 699 | 1.7% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 150 | 0.4% | |
| 1.0 | 3730 | 9.0% | |
| 2.0 | 2952 | 7.1% | |
| 3.0 | 3602 | 8.6% | |
| 4.0 | 2917 | 7.0% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 9.0 | 3175 | 7.6% | |
| 10.0 | 475 | 1.1% | |
| 11.0 | 73 | 0.2% | |
| 13.0 | 1 | 0.0% | |
| 99.0 | 1270 | 3.0% |
Dual
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.1204 |
|---|
| 0 | |
|---|---|
| 1 | 5017 |
| Value | Count | Frequency (%) | |
| 0 | 36652 | 88.0% | |
| 1 | 5017 | 12.0% |
Elevation
Highly correlated
This variable is highly correlated with Slope and should be ignored for analysis
| Correlation | 0.99764 |
|---|
FEMA
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.054237 |
|---|
| 0 | |
|---|---|
| 1 | 2260 |
| Value | Count | Frequency (%) | |
| 0 | 39409 | 94.6% | |
| 1 | 2260 | 5.4% |
FO_Land_Owner
Numeric
| Distinct count | 22 |
|---|---|
| Unique (%) | 0.1% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 18.305 |
|---|---|
| Minimum | 0 |
| Maximum | 90 |
| Zeros (%) | 0.1% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 10 |
| Q1 | 10 |
| Median | 11 |
| Q3 | 20 |
| 95-th percentile | 50 |
| Maximum | 90 |
| Range | 90 |
| Interquartile range | 10 |
Descriptive statistics
| Standard deviation | 15.955 |
|---|---|
| Coef of variation | 0.87165 |
| Kurtosis | 5.0835 |
| Mean | 18.305 |
| MAD | 11.361 |
| Skewness | 2.2726 |
| Sum | 762740 |
| Variance | 254.57 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 10.0 | 19695 | 47.3% | |
| 13.0 | 4951 | 11.9% | |
| 50.0 | 4444 | 10.7% | |
| 11.0 | 3978 | 9.5% | |
| 20.0 | 2959 | 7.1% | |
| 12.0 | 2047 | 4.9% | |
| 30.0 | 1540 | 3.7% | |
| 40.0 | 734 | 1.8% | |
| 90.0 | 573 | 1.4% | |
| 70.0 | 194 | 0.5% | |
| Other values (12) | 554 | 1.3% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 59 | 0.1% | |
| 1.0 | 172 | 0.4% | |
| 2.0 | 24 | 0.1% | |
| 3.0 | 9 | 0.0% | |
| 4.0 | 3 | 0.0% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 50.0 | 4444 | 10.7% | |
| 60.0 | 118 | 0.3% | |
| 70.0 | 194 | 0.5% | |
| 80.0 | 9 | 0.0% | |
| 90.0 | 573 | 1.4% |
Federal_lands
Numeric
| Distinct count | 4 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 1.0473 |
|---|---|
| Minimum | 0 |
| Maximum | 3 |
| Zeros (%) | 8.3% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 0 |
| Q1 | 1 |
| Median | 1 |
| Q3 | 1 |
| 95-th percentile | 3 |
| Maximum | 3 |
| Range | 3 |
| Interquartile range | 0 |
Descriptive statistics
| Standard deviation | 0.56382 |
|---|---|
| Coef of variation | 0.53838 |
| Kurtosis | 5.8411 |
| Mean | 1.0473 |
| MAD | 0.2537 |
| Skewness | 1.792 |
| Sum | 43638 |
| Variance | 0.31789 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 1.0 | 34978 | 83.9% | |
| 0.0 | 3469 | 8.3% | |
| 3.0 | 2216 | 5.3% | |
| 2.0 | 1006 | 2.4% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 3469 | 8.3% | |
| 1.0 | 34978 | 83.9% | |
| 2.0 | 1006 | 2.4% | |
| 3.0 | 2216 | 5.3% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 3469 | 8.3% | |
| 1.0 | 34978 | 83.9% | |
| 2.0 | 1006 | 2.4% | |
| 3.0 | 2216 | 5.3% |
FireCategory
Constant
This variable is constant and should be ignored for analysis
| Constant value | 1 |
|---|
FireYear
Numeric
| Distinct count | 59 |
|---|---|
| Unique (%) | 0.1% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 1992 |
|---|---|
| Minimum | 1960 |
| Maximum | 2019 |
| Zeros (%) | 0.0% |
Quantile statistics
| Minimum | 1960 |
|---|---|
| 5-th percentile | 1969 |
| Q1 | 1979 |
| Median | 1992 |
| Q3 | 2005 |
| 95-th percentile | 2016 |
| Maximum | 2019 |
| Range | 59 |
| Interquartile range | 26 |
Descriptive statistics
| Standard deviation | 15.017 |
|---|---|
| Coef of variation | 0.0075386 |
| Kurtosis | -1.1464 |
| Mean | 1992 |
| MAD | 12.864 |
| Skewness | 0.064528 |
| Sum | 83005611 |
| Variance | 225.51 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 1987 | 1249 | 3.0% | |
| 1970 | 1219 | 2.9% | |
| 1992 | 1207 | 2.9% | |
| 1994 | 1043 | 2.5% | |
| 1974 | 1014 | 2.4% | |
| 1975 | 995 | 2.4% | |
| 1973 | 986 | 2.4% | |
| 2006 | 943 | 2.3% | |
| 2001 | 934 | 2.2% | |
| 1985 | 920 | 2.2% | |
| Other values (49) | 31159 | 74.8% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 1960 | 4 | 0.0% | |
| 1961 | 6 | 0.0% | |
| 1962 | 2 | 0.0% | |
| 1963 | 2 | 0.0% | |
| 1964 | 5 | 0.0% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 2015 | 771 | 1.9% | |
| 2016 | 594 | 1.4% | |
| 2017 | 812 | 1.9% | |
| 2018 | 820 | 2.0% | |
| 2019 | 408 | 1.0% |
Flame_length
Numeric
| Distinct count | 7 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 0.40308 |
|---|---|
| Minimum | 0 |
| Maximum | 6 |
| Zeros (%) | 70.8% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 0 |
| Q1 | 0 |
| Median | 0 |
| Q3 | 1 |
| 95-th percentile | 2 |
| Maximum | 6 |
| Range | 6 |
| Interquartile range | 1 |
Descriptive statistics
| Standard deviation | 0.77548 |
|---|---|
| Coef of variation | 1.9239 |
| Kurtosis | 12.563 |
| Mean | 0.40308 |
| MAD | 0.57116 |
| Skewness | 2.9311 |
| Sum | 16796 |
| Variance | 0.60137 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 0.0 | 29522 | 70.8% | |
| 1.0 | 9156 | 22.0% | |
| 2.0 | 2068 | 5.0% | |
| 3.0 | 530 | 1.3% | |
| 4.0 | 183 | 0.4% | |
| 6.0 | 132 | 0.3% | |
| 5.0 | 78 | 0.2% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 29522 | 70.8% | |
| 1.0 | 9156 | 22.0% | |
| 2.0 | 2068 | 5.0% | |
| 3.0 | 530 | 1.3% | |
| 4.0 | 183 | 0.4% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 2.0 | 2068 | 5.0% | |
| 3.0 | 530 | 1.3% | |
| 4.0 | 183 | 0.4% | |
| 5.0 | 78 | 0.2% | |
| 6.0 | 132 | 0.3% |
Fuel
Numeric
| Distinct count | 15 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 15.588 |
|---|---|
| Minimum | 0 |
| Maximum | 99 |
| Zeros (%) | 0.4% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 1 |
| Q1 | 3 |
| Median | 5 |
| Q3 | 9 |
| 95-th percentile | 99 |
| Maximum | 99 |
| Range | 99 |
| Interquartile range | 6 |
Descriptive statistics
| Standard deviation | 29.31 |
|---|---|
| Coef of variation | 1.8803 |
| Kurtosis | 4.1519 |
| Mean | 15.588 |
| MAD | 18.132 |
| Skewness | 2.456 |
| Sum | 649520 |
| Variance | 859.06 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 3.0 | 7048 | 16.9% | |
| 5.0 | 5433 | 13.0% | |
| 1.0 | 4938 | 11.9% | |
| 8.0 | 4919 | 11.8% | |
| 99.0 | 4529 | 10.9% | |
| 2.0 | 3689 | 8.9% | |
| 7.0 | 2590 | 6.2% | |
| 12.0 | 2432 | 5.8% | |
| 9.0 | 1920 | 4.6% | |
| 6.0 | 1276 | 3.1% | |
| Other values (5) | 2895 | 6.9% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 148 | 0.4% | |
| 1.0 | 4938 | 11.9% | |
| 2.0 | 3689 | 8.9% | |
| 3.0 | 7048 | 16.9% | |
| 4.0 | 214 | 0.5% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 10.0 | 1255 | 3.0% | |
| 11.0 | 549 | 1.3% | |
| 12.0 | 2432 | 5.8% | |
| 13.0 | 729 | 1.7% | |
| 99.0 | 4529 | 10.9% |
General
Numeric
| Distinct count | 11 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 3.8775 |
|---|---|
| Minimum | 0 |
| Maximum | 10 |
| Zeros (%) | 0.0% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 1 |
| Q1 | 1 |
| Median | 4 |
| Q3 | 6 |
| 95-th percentile | 9 |
| Maximum | 10 |
| Range | 10 |
| Interquartile range | 5 |
Descriptive statistics
| Standard deviation | 2.5849 |
|---|---|
| Coef of variation | 0.66663 |
| Kurtosis | -0.91615 |
| Mean | 3.8775 |
| MAD | 2.2163 |
| Skewness | 0.43665 |
| Sum | 161570 |
| Variance | 6.6816 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 1.0 | 13587 | 32.6% | |
| 6.0 | 7819 | 18.8% | |
| 3.0 | 6375 | 15.3% | |
| 4.0 | 4245 | 10.2% | |
| 5.0 | 3292 | 7.9% | |
| 9.0 | 3075 | 7.4% | |
| 8.0 | 1610 | 3.9% | |
| 7.0 | 802 | 1.9% | |
| 2.0 | 784 | 1.9% | |
| 10.0 | 77 | 0.2% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 3 | 0.0% | |
| 1.0 | 13587 | 32.6% | |
| 2.0 | 784 | 1.9% | |
| 3.0 | 6375 | 15.3% | |
| 4.0 | 4245 | 10.2% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 6.0 | 7819 | 18.8% | |
| 7.0 | 802 | 1.9% | |
| 8.0 | 1610 | 3.9% | |
| 9.0 | 3075 | 7.4% | |
| 10.0 | 77 | 0.2% |
Land_Class
Numeric
| Distinct count | 12 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 1.769 |
|---|---|
| Minimum | 0 |
| Maximum | 11 |
| Zeros (%) | 0.7% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 1 |
| Q1 | 1 |
| Median | 1 |
| Q3 | 1 |
| 95-th percentile | 7 |
| Maximum | 11 |
| Range | 11 |
| Interquartile range | 0 |
Descriptive statistics
| Standard deviation | 1.8189 |
|---|---|
| Coef of variation | 1.0282 |
| Kurtosis | 5.0591 |
| Mean | 1.769 |
| MAD | 1.1956 |
| Skewness | 2.473 |
| Sum | 73713 |
| Variance | 3.3083 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 1.0 | 31710 | 76.1% | |
| 2.0 | 4072 | 9.8% | |
| 7.0 | 1524 | 3.7% | |
| 5.0 | 1170 | 2.8% | |
| 6.0 | 748 | 1.8% | |
| 4.0 | 747 | 1.8% | |
| 8.0 | 744 | 1.8% | |
| 3.0 | 350 | 0.8% | |
| 0.0 | 296 | 0.7% | |
| 9.0 | 218 | 0.5% | |
| Other values (2) | 90 | 0.2% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 296 | 0.7% | |
| 1.0 | 31710 | 76.1% | |
| 2.0 | 4072 | 9.8% | |
| 3.0 | 350 | 0.8% | |
| 4.0 | 747 | 1.8% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 7.0 | 1524 | 3.7% | |
| 8.0 | 744 | 1.8% | |
| 9.0 | 218 | 0.5% | |
| 10.0 | 89 | 0.2% | |
| 11.0 | 1 | 0.0% |
Land_Class_tg
Numeric
| Distinct count | 3 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 0.57018 |
|---|---|
| Minimum | 0 |
| Maximum | 2 |
| Zeros (%) | 49.9% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 0 |
| Q1 | 0 |
| Median | 1 |
| Q3 | 1 |
| 95-th percentile | 2 |
| Maximum | 2 |
| Range | 2 |
| Interquartile range | 1 |
Descriptive statistics
| Standard deviation | 0.61843 |
|---|---|
| Coef of variation | 1.0846 |
| Kurtosis | -0.57603 |
| Mean | 0.57018 |
| MAD | 0.56847 |
| Skewness | 0.6035 |
| Sum | 23759 |
| Variance | 0.38245 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 0.0 | 20772 | 49.9% | |
| 1.0 | 18035 | 43.3% | |
| 2.0 | 2862 | 6.9% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 20772 | 49.9% | |
| 1.0 | 18035 | 43.3% | |
| 2.0 | 2862 | 6.9% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 20772 | 49.9% | |
| 1.0 | 18035 | 43.3% | |
| 2.0 | 2862 | 6.9% |
Latitude
Highly correlated
This variable is highly correlated with Y and should be ignored for analysis
| Correlation | 0.99991 |
|---|
Longitude
Highly correlated
This variable is highly correlated with X and should be ignored for analysis
| Correlation | 0.99964 |
|---|
Minimum
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.12388 |
|---|
| 0 | |
|---|---|
| 1 | 5162 |
| Value | Count | Frequency (%) | |
| 0 | 36507 | 87.6% | |
| 1 | 5162 | 12.4% |
Protection_agency
Numeric
| Distinct count | 10 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 1.1153 |
|---|---|
| Minimum | 0 |
| Maximum | 9 |
| Zeros (%) | 0.3% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 1 |
| Q1 | 1 |
| Median | 1 |
| Q3 | 1 |
| 95-th percentile | 1 |
| Maximum | 9 |
| Range | 9 |
| Interquartile range | 0 |
Descriptive statistics
| Standard deviation | 0.73505 |
|---|---|
| Coef of variation | 0.65905 |
| Kurtosis | 44.797 |
| Mean | 1.1153 |
| MAD | 0.23057 |
| Skewness | 6.5649 |
| Sum | 46474 |
| Variance | 0.54029 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 1.0 | 40286 | 96.7% | |
| 6.0 | 491 | 1.2% | |
| 3.0 | 314 | 0.8% | |
| 5.0 | 157 | 0.4% | |
| 0.0 | 142 | 0.3% | |
| 7.0 | 107 | 0.3% | |
| 4.0 | 91 | 0.2% | |
| 2.0 | 46 | 0.1% | |
| 9.0 | 30 | 0.1% | |
| 8.0 | 5 | 0.0% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 142 | 0.3% | |
| 1.0 | 40286 | 96.7% | |
| 2.0 | 46 | 0.1% | |
| 3.0 | 314 | 0.8% | |
| 4.0 | 91 | 0.2% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 5.0 | 157 | 0.4% | |
| 6.0 | 491 | 1.2% | |
| 7.0 | 107 | 0.3% | |
| 8.0 | 5 | 0.0% | |
| 9.0 | 30 | 0.1% |
SB_360_Liable
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.0094795 |
|---|
| 0.0 | |
|---|---|
| 1.0 | 395 |
| Value | Count | Frequency (%) | |
| 0.0 | 41274 | 99.1% | |
| 1.0 | 395 | 0.9% |
SB_360_Lot
Numeric
| Distinct count | 3 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 0.13478 |
|---|---|
| Minimum | 0 |
| Maximum | 2 |
| Zeros (%) | 92.0% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 0 |
| Q1 | 0 |
| Median | 0 |
| Q3 | 0 |
| 95-th percentile | 2 |
| Maximum | 2 |
| Range | 2 |
| Interquartile range | 0 |
Descriptive statistics
| Standard deviation | 0.47641 |
|---|---|
| Coef of variation | 3.5348 |
| Kurtosis | 10.238 |
| Mean | 0.13478 |
| MAD | 0.2481 |
| Skewness | 3.437 |
| Sum | 5616 |
| Variance | 0.22696 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 0.0 | 38352 | 92.0% | |
| 2.0 | 2299 | 5.5% | |
| 1.0 | 1018 | 2.4% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 38352 | 92.0% | |
| 1.0 | 1018 | 2.4% | |
| 2.0 | 2299 | 5.5% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 38352 | 92.0% | |
| 1.0 | 1018 | 2.4% | |
| 2.0 | 2299 | 5.5% |
Size_at_attack
Numeric
| Distinct count | 285 |
|---|---|
| Unique (%) | 0.7% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 3.0355 |
|---|---|
| Minimum | 0 |
| Maximum | 50815 |
| Zeros (%) | 51.8% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 0 |
| Q1 | 0 |
| Median | 0 |
| Q3 | 0.1 |
| 95-th percentile | 1.5 |
| Maximum | 50815 |
| Range | 50815 |
| Interquartile range | 0.1 |
Descriptive statistics
| Standard deviation | 260.28 |
|---|---|
| Coef of variation | 85.748 |
| Kurtosis | 34968 |
| Mean | 3.0355 |
| MAD | 5.6317 |
| Skewness | 181.37 |
| Sum | 126480 |
| Variance | 67748 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 0.0 | 21569 | 51.8% | |
| 0.01 | 7063 | 17.0% | |
| 0.1 | 4409 | 10.6% | |
| 0.25 | 1893 | 4.5% | |
| 0.5 | 1022 | 2.5% | |
| 1.0 | 793 | 1.9% | |
| 0.2 | 539 | 1.3% | |
| 2.0 | 427 | 1.0% | |
| 0.02 | 409 | 1.0% | |
| 1.5 | 290 | 0.7% | |
| Other values (275) | 3255 | 7.8% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 21569 | 51.8% | |
| 0.01 | 7063 | 17.0% | |
| 0.02 | 409 | 1.0% | |
| 0.03 | 110 | 0.3% | |
| 0.04 | 48 | 0.1% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 3000.0 | 3 | 0.0% | |
| 3723.0 | 1 | 0.0% | |
| 7050.0 | 1 | 0.0% | |
| 11615.0 | 1 | 0.0% | |
| 50815.0 | 1 | 0.0% |
Slope
Highly correlated
This variable is highly correlated with Aspect and should be ignored for analysis
| Correlation | 0.99941 |
|---|
SurchargeLot
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.10295 |
|---|
| 0 | |
|---|---|
| 1 | 4290 |
| Value | Count | Frequency (%) | |
| 0 | 37379 | 89.7% | |
| 1 | 4290 | 10.3% |
SurchargeLotAssessed
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.066812 |
|---|
| 0 | |
|---|---|
| 1 | 2784 |
| Value | Count | Frequency (%) | |
| 0 | 38885 | 93.3% | |
| 1 | 2784 | 6.7% |
Topography
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.33677 |
|---|
| 0 | |
|---|---|
| 1 |
| Value | Count | Frequency (%) | |
| 0 | 27636 | 66.3% | |
| 1 | 14033 | 33.7% |
X
Numeric
| Distinct count | 31863 |
|---|---|
| Unique (%) | 76.5% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 881940 |
|---|---|
| Minimum | 224480 |
| Maximum | 2330100 |
| Zeros (%) | 0.0% |
Quantile statistics
| Minimum | 224480 |
|---|---|
| 5-th percentile | 367460 |
| Q1 | 545730 |
| Median | 705170 |
| Q3 | 1074300 |
| 95-th percentile | 1936900 |
| Maximum | 2330100 |
| Range | 2105600 |
| Interquartile range | 528560 |
Descriptive statistics
| Standard deviation | 469600 |
|---|---|
| Coef of variation | 0.53246 |
| Kurtosis | 0.36579 |
| Mean | 881940 |
| MAD | 375690 |
| Skewness | 1.1456 |
| Sum | 36749000000 |
| Variance | 220520000000 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 526103.824146986 | 28 | 0.1% | |
| 562406.384842515 | 25 | 0.1% | |
| 536428.536745414 | 24 | 0.1% | |
| 571746.639435694 | 22 | 0.1% | |
| 546578.327099741 | 21 | 0.1% | |
| 457387.64238844794 | 20 | 0.0% | |
| 428648.02460630203 | 19 | 0.0% | |
| 545210.855314955 | 19 | 0.0% | |
| 623516.884514436 | 18 | 0.0% | |
| 536614.990157485 | 17 | 0.0% | |
| Other values (31853) | 41456 | 99.5% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 224481.370734915 | 2 | 0.0% | |
| 226531.344816267 | 1 | 0.0% | |
| 228714.39796587802 | 1 | 0.0% | |
| 229226.158792645 | 1 | 0.0% | |
| 231490.077427819 | 1 | 0.0% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 2270734.09120736 | 1 | 0.0% | |
| 2271832.16240157 | 1 | 0.0% | |
| 2288277.34776902 | 1 | 0.0% | |
| 2293224.94717848 | 1 | 0.0% | |
| 2330062.90485564 | 1 | 0.0% |
Y
Numeric
| Distinct count | 31863 |
|---|---|
| Unique (%) | 76.5% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 743900 |
|---|---|
| Minimum | 54599 |
| Maximum | 1656700 |
| Zeros (%) | 0.0% |
Quantile statistics
| Minimum | 54599 |
|---|---|
| 5-th percentile | 155730 |
| Q1 | 295660 |
| Median | 701500 |
| Q3 | 1144000 |
| 95-th percentile | 1478800 |
| Maximum | 1656700 |
| Range | 1602100 |
| Interquartile range | 848320 |
Descriptive statistics
| Standard deviation | 454330 |
|---|---|
| Coef of variation | 0.61075 |
| Kurtosis | -1.2942 |
| Mean | 743900 |
| MAD | 403300 |
| Skewness | 0.25417 |
| Sum | 30997000000 |
| Variance | 206420000000 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 292139.513779521 | 28 | 0.1% | |
| 264508.78083989 | 25 | 0.1% | |
| 286449.714566931 | 24 | 0.1% | |
| 696971.049212605 | 22 | 0.1% | |
| 275555.029855639 | 21 | 0.1% | |
| 161345.489501312 | 20 | 0.0% | |
| 228183.73392388198 | 19 | 0.0% | |
| 1643614.27559055 | 19 | 0.0% | |
| 187974.796259835 | 18 | 0.0% | |
| 297458.22473754 | 17 | 0.0% | |
| Other values (31853) | 41456 | 99.5% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 54598.510826766505 | 1 | 0.0% | |
| 66764.41568242009 | 1 | 0.0% | |
| 90169.05479002 | 1 | 0.0% | |
| 90816.4665354341 | 1 | 0.0% | |
| 90865.3143044561 | 1 | 0.0% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 1649814.66699475 | 1 | 0.0% | |
| 1651909.77132545 | 1 | 0.0% | |
| 1652731.50098425 | 1 | 0.0% | |
| 1652765.31102362 | 3 | 0.0% | |
| 1656677.58923885 | 1 | 0.0% |
cluster
Numeric
| Distinct count | 10 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 4.125 |
|---|---|
| Minimum | 0 |
| Maximum | 9 |
| Zeros (%) | 11.6% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 0 |
| Q1 | 2 |
| Median | 4 |
| Q3 | 7 |
| 95-th percentile | 9 |
| Maximum | 9 |
| Range | 9 |
| Interquartile range | 5 |
Descriptive statistics
| Standard deviation | 2.803 |
|---|---|
| Coef of variation | 0.6795 |
| Kurtosis | -1.065 |
| Mean | 4.125 |
| MAD | 2.3494 |
| Skewness | 0.23585 |
| Sum | 171886 |
| Variance | 7.8566 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 4 | 6562 | 15.7% | |
| 2 | 5349 | 12.8% | |
| 0 | 4815 | 11.6% | |
| 3 | 4707 | 11.3% | |
| 1 | 3713 | 8.9% | |
| 9 | 3622 | 8.7% | |
| 8 | 3587 | 8.6% | |
| 5 | 3311 | 7.9% | |
| 7 | 3239 | 7.8% | |
| 6 | 2764 | 6.6% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0 | 4815 | 11.6% | |
| 1 | 3713 | 8.9% | |
| 2 | 5349 | 12.8% | |
| 3 | 4707 | 11.3% | |
| 4 | 6562 | 15.7% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 5 | 3311 | 7.9% | |
| 6 | 2764 | 6.6% | |
| 7 | 3239 | 7.8% | |
| 8 | 3587 | 8.6% | |
| 9 | 3622 | 8.7% |
cluster_0
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.11555 |
|---|
| 0 | |
|---|---|
| 1 | 4815 |
| Value | Count | Frequency (%) | |
| 0 | 36854 | 88.4% | |
| 1 | 4815 | 11.6% |
cluster_1
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.089107 |
|---|
| 0 | |
|---|---|
| 1 | 3713 |
| Value | Count | Frequency (%) | |
| 0 | 37956 | 91.1% | |
| 1 | 3713 | 8.9% |
cluster_2
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.12837 |
|---|
| 0 | |
|---|---|
| 1 | 5349 |
| Value | Count | Frequency (%) | |
| 0 | 36320 | 87.2% | |
| 1 | 5349 | 12.8% |
cluster_3
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.11296 |
|---|
| 0 | |
|---|---|
| 1 | 4707 |
| Value | Count | Frequency (%) | |
| 0 | 36962 | 88.7% | |
| 1 | 4707 | 11.3% |
cluster_4
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.15748 |
|---|
| 0 | |
|---|---|
| 1 | 6562 |
| Value | Count | Frequency (%) | |
| 0 | 35107 | 84.3% | |
| 1 | 6562 | 15.7% |
cluster_5
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.07946 |
|---|
| 0 | |
|---|---|
| 1 | 3311 |
| Value | Count | Frequency (%) | |
| 0 | 38358 | 92.1% | |
| 1 | 3311 | 7.9% |
cluster_6
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.066332 |
|---|
| 0 | |
|---|---|
| 1 | 2764 |
| Value | Count | Frequency (%) | |
| 0 | 38905 | 93.4% | |
| 1 | 2764 | 6.6% |
cluster_7
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.077732 |
|---|
| 0 | |
|---|---|
| 1 | 3239 |
| Value | Count | Frequency (%) | |
| 0 | 38430 | 92.2% | |
| 1 | 3239 | 7.8% |
cluster_8
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.086083 |
|---|
| 0 | |
|---|---|
| 1 | 3587 |
| Value | Count | Frequency (%) | |
| 0 | 38082 | 91.4% | |
| 1 | 3587 | 8.6% |
cluster_9
Boolean
| Distinct count | 2 |
|---|---|
| Unique (%) | 0.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Mean | 0.086923 |
|---|
| 0 | |
|---|---|
| 1 | 3622 |
| Value | Count | Frequency (%) | |
| 0 | 38047 | 91.3% | |
| 1 | 3622 | 8.7% |
fire_dayofyear
Numeric
| Distinct count | 362 |
|---|---|
| Unique (%) | 0.9% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 211.73 |
|---|---|
| Minimum | 0 |
| Maximum | 363 |
| Zeros (%) | 0.4% |
Quantile statistics
| Minimum | 0 |
|---|---|
| 5-th percentile | 129 |
| Q1 | 186 |
| Median | 214 |
| Q3 | 242 |
| 95-th percentile | 285 |
| Maximum | 363 |
| Range | 363 |
| Interquartile range | 56 |
Descriptive statistics
| Standard deviation | 48.275 |
|---|---|
| Coef of variation | 0.22801 |
| Kurtosis | 1.8562 |
| Mean | 211.73 |
| MAD | 36.111 |
| Skewness | -0.72874 |
| Sum | 8822400 |
| Variance | 2330.5 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 207.0 | 593 | 1.4% | |
| 213.0 | 570 | 1.4% | |
| 222.0 | 546 | 1.3% | |
| 221.0 | 471 | 1.1% | |
| 220.0 | 469 | 1.1% | |
| 212.0 | 469 | 1.1% | |
| 217.0 | 452 | 1.1% | |
| 208.0 | 449 | 1.1% | |
| 224.0 | 445 | 1.1% | |
| 219.0 | 442 | 1.1% | |
| Other values (352) | 36763 | 88.2% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 0.0 | 162 | 0.4% | |
| 2.0 | 3 | 0.0% | |
| 3.0 | 1 | 0.0% | |
| 4.0 | 1 | 0.0% | |
| 5.0 | 5 | 0.0% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 359.0 | 3 | 0.0% | |
| 360.0 | 2 | 0.0% | |
| 361.0 | 3 | 0.0% | |
| 362.0 | 2 | 0.0% | |
| 363.0 | 1 | 0.0% |
index
Numeric
| Distinct count | 41669 |
|---|---|
| Unique (%) | 100.0% |
| Missing (%) | 0.0% |
| Missing (n) | 0 |
| Infinite (%) | 0.0% |
| Infinite (n) | 0 |
| Mean | 31015 |
|---|---|
| Minimum | 1 |
| Maximum | 64052 |
| Zeros (%) | 0.0% |
Quantile statistics
| Minimum | 1 |
|---|---|
| 5-th percentile | 2877.4 |
| Q1 | 14101 |
| Median | 28361 |
| Q3 | 49577 |
| 95-th percentile | 61217 |
| Maximum | 64052 |
| Range | 64051 |
| Interquartile range | 35476 |
Descriptive statistics
| Standard deviation | 19380 |
|---|---|
| Coef of variation | 0.62488 |
| Kurtosis | -1.3213 |
| Mean | 31015 |
| MAD | 17010 |
| Skewness | 0.13852 |
| Sum | 1292347969 |
| Variance | 375590000 |
| Memory size | 325.6 KiB |
| Value | Count | Frequency (%) | |
| 2047 | 1 | 0.0% | |
| 48541 | 1 | 0.0% | |
| 52659 | 1 | 0.0% | |
| 50610 | 1 | 0.0% | |
| 56753 | 1 | 0.0% | |
| 54704 | 1 | 0.0% | |
| 9646 | 1 | 0.0% | |
| 15789 | 1 | 0.0% | |
| 13740 | 1 | 0.0% | |
| 3499 | 1 | 0.0% | |
| Other values (41659) | 41659 | 100.0% |
Minimum 5 values
| Value | Count | Frequency (%) | |
| 1 | 1 | 0.0% | |
| 2 | 1 | 0.0% | |
| 3 | 1 | 0.0% | |
| 4 | 1 | 0.0% | |
| 9 | 1 | 0.0% |
Maximum 5 values
| Value | Count | Frequency (%) | |
| 64046 | 1 | 0.0% | |
| 64048 | 1 | 0.0% | |
| 64049 | 1 | 0.0% | |
| 64050 | 1 | 0.0% | |
| 64052 | 1 | 0.0% |
Correlations
Sample
In [60]:
# Calculating Baseline Prediction errors
from sklearn.metrics import mean_absolute_error as MAE
from sklearn.metrics import mean_squared_error as MSE
# Base prediction is mean of the y_train
y_base = [np.mean(y_train)]*y_test.shape[0]
In [61]:
# Mean Errors for Baseline Prediction
prediction_errors = {}
prediction_errors['Base'] = [MAE(y_test,y_base),MSE(y_test,y_base)]
# prediction_errors['Base MSE'] = MAE(y_test,y_base)
print(f'Mean absolute error for baseline prediction is {MAE(y_test,y_base)}')
print(f'Mean standard error for baseline prediction is {MSE(y_test,y_base)}')
In [62]:
# Using Logistic Regression to train the model
from sklearn.linear_model import LinearRegression
lr = LinearRegression()
lr.fit(X_train, y_train)
Out[62]:
In [63]:
# MAE and MSE using Linear Regression
y_pred = lr.predict(X_test)
# Errors for Linear Prediction
prediction_errors['Linear Regression'] = [MAE(y_test,y_pred),MSE(y_test,y_pred)]
print(f'Mean absolute error for Linear Regression is {MAE(y_test,y_pred)}')
print(f'Mean standard error for Linear Regression is {MSE(y_test,y_pred)}')
In [64]:
# Using RandomForest to train the model
from sklearn.ensemble import RandomForestRegressor
rf = RandomForestRegressor(n_estimators=100)
rf.fit(X_train, y_train)
Out[64]:
In [65]:
# MAE and MSE using Random Forest Regression
y_pred = rf.predict(X_test)
# Mean Errors for Random Forest Prediction
prediction_errors['Random Forest'] = [MAE(y_test,y_pred),MSE(y_test,y_pred)]
print(f'Mean absolute error for Linear Regression is {MAE(y_test,y_pred)}')
print(f'Mean standard error for Linear Regression is {MSE(y_test,y_pred)}')
In [66]:
# Training the data using Ridge Regression
from sklearn.linear_model import Ridge
rr = Ridge(alpha=1.0)
rr.fit(X_train, y_train)
Out[66]:
In [68]:
# MAE and MSE using Ridge
y_pred = rr.predict(X_test)
# Mean Errors for Ridge Prediction
prediction_errors['Ridge Regression'] = [MAE(y_test,y_pred),MSE(y_test,y_pred)]
print(f'Mean absolute error for Ridge is {MAE(y_test,y_pred)}')
print(f'Mean standard error for Ridge is {MSE(y_test,y_pred)}')
In [71]:
# Building a Neural network to predict the area burned by fire in Oregon
model = Sequential()
model.add(Dense(35, input_dim=n, kernel_initializer='normal', activation='relu'))
model.add(Dense(70, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(256, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(56, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(28, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(14, activation='relu'))
model.add(Dense(8, activation='relu'))
model.add(Dense(1, activation='linear'))
model.summary()
In [72]:
model.compile(loss='mae', optimizer='adam', metrics=['mse','mae'])
In [73]:
history = model.fit(X_train, y_train.values, epochs=20, batch_size=1028, verbose = 2, validation_split = 0.05, shuffle=True)
In [74]:
# MAE and MSE using Neural Networks
y_pred = model.predict(X_test)
# Mean Errors for Neural Networks
prediction_errors['Neural Networks'] = [MAE(y_test,y_pred),MSE(y_test,y_pred)]
print(f'Mean absolute error for NN is {MAE(y_test,y_pred)}')
print(f'Mean standard error for NN is {MSE(y_test,y_pred)}')
In [75]:
# Plot training & validation loss values for Neural Networks
plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model loss')
plt.ylabel('Loss: Mean Absolue Error')
plt.xlabel('Epoch')
plt.legend(['Train', 'Val'], loc='upper right')
plt.show()
Out[75]:
In [77]:
# Summary of errors for all models used
prediction_errors = pd.DataFrame(prediction_errors)
prediction_errors.index = ['Mean Absoluter Error', 'Mean Standard Error']
prediction_errors.T
Out[77]: